How to Win a Euro Referendum: Understanding Mass Support for British Membership of the Single Currency

How to Win a Euro Referendum: Understanding Mass Support for British Membership of the Single Currency Matthew Gabel University of Kentucky Simon Hix London School of Economics and Political Science Draft 11 November 2002 Please do not cite

Abstract With the possibility of a referendum on British membership of the Euro some time in 2003 or 2004, the relevance of British public opinion on the Euro is obvious. There has been much public debate about whether Tony Blair could win a referendum despite opinion polls showing a widening gap between those opposed to membership and those in favor. However, there has been little academic research on what explains mass British attitudes towards the Euro at an individual level, and hence what factors might be crucial in a referendum campaign. Some research has highlighted the importance of British citizens sense of national identity. In this paper we explore these and other explanations. Specifically, we examine whether one s assessment of EU membership and the quality of EU governing institutions affects one s support for the Euro. We find national identity affects support for a single currency, but it is hardly the only or most important factor. Voters levels of information about the Euro, their concerns over the democratic character of EU governance and their perceptions of the benefits of EU membership are also important determinants of support. 2

Introduction Whether Britain should join the European single currency, the Euro, was a central issue in the 2001 British general election. On one side, the Conservative Party ran a clearly anti-euro campaign, promising to save the Pound. On the other side, the Labour Party promised to make an assessment of the economic benefits of British membership of the Euro, before putting the issue to the public in a referendum. The British Chancellor of the Exchequer (Finance Minister), Gordon Brown, subsequently defined a series of economic tests, and has promised to make a decision in June 2003 about whether Britain meets these tests, and hence whether he will recommend British membership of the Euro. However, for Tony Blair, the British Prime Minister, the issue is more than simply about economics. British membership of the Euro is central to Britain s political commitment to the project of European integration, is the single most important issue of his premiership, and is the legacy Blair wants to be remembered by (akin to Chancellor Helmut Kohl s role in German unification!). If in June 2003 Brown concludes that the economic conditions for British membership are favorable, the next question for the Labour leadership will be when to hold a referendum, and what strategy they should take to maximize their chances of securing a positive vote. The problem for Blair and Brown is that in the last decade, public opinion surveys have consistently indicated that a referendum on British adoption of the Euro could result in a negative vote. For example, Eurobarometer surveys since 1992 have never shown more than 40 percent of the British public in favor of a common currency (Hix 2000; Banducci, Karp, and Loedel 2002). Moreover, recent Eurobarometer surveys indicate declining support, with less than a third supporting the Euro, and a clear majority opposing membership. Surveys using different questions or conducted by different organizations generally confirm this description of mass British attitudes toward the Euro. 3

Given this level of opposition, one might simply conclude that it would be very difficult for Blair and Brown to win a Euro referendum, and hence that a referendum on the issue is unlikely to be held in the near future. However, there are at least two reasons to question this assessment. First, the same surveys showing public opposition to the Euro also indicate that many British citizens consider themselves poorly informed about the Euro and, as a result, express ambivalence about the issue. Thus, the aggregate trend in opposition to the Euro may not provide a clear prediction of individual level voting behavior in a referendum, particularly if it were preceded by an informative campaign on both sides (Whitely 2001; Atkinson 2001). Indeed, as the 1975 British referendum on membership in the European Community showed, public opposition can decline abruptly in the midst of a strong campaign in support of the issue by business and government (Dalton and Duval 1986). Consequently, to forecast voting behavior in a Euro referendum, we need to understand why British citizens differ in their support at an individual level, not just the aggregate level of support in an opinion survey. Second, a well-funded and vocal campaign on both sides will very likely accompany any British Euro referendum. Thus, for those interested in changing opinion toward the Euro, an understanding of why British differ in their support for the Euro would help craft a campaign strategy. On this point, scholars and journalists already speculate on how and why aspects of the current no campaign have been effective in building opposition to the Euro (Leonard and Arbuthnott 2001). But we do not yet have much analysis about whether such speculations are correct and thus what types of strategies might be valuable for altering public opinion. In sum, our expectations about the outcome of a referendum on the Euro and our ability to strategize effectively to influence such an outcome depend on our understanding of why British citizens 4

differ in their support for the Euro, what factors shape these differences, and how malleable such factors are to campaign strategies. What factors account for variation in British citizens support for the Euro? Previous research offers three types of answers. First, several studies have developed theories with implications for individual-level variation in European Union (EU) citizens support for the Euro (Scheve 1999; Gabel 2001; Banducci, Karp, and Loedel 2002; M ller-peters 1998). When applied to the British case, these general theories generate inferences about which types of citizens are likely to support the Euro and why. For example, Scheve (1999) developed a theoretical model predicting that support for the Euro rises with pre-tax income. Second, many of these same studies identify empirical regularities in support for the Euro. For instance, Gabel (2001) shows that voters on the Left of the Left-Right dimension are less supportive of a single currency than voters of the Right. If these regularities apply to the British case, then they would identify sources of support and opposition. Third, Routh and Bugonye (1998) provide evidence regarding how national identity relates to support for the Euro among British citizens. While these three types of studies all provide potential insights into why British citizens vary in their support for the Euro, they all have potential problems. As M ller-peters (1998) concedes, British attitudes toward the Euro appear to be structured differently than those of citizens in other EU member states. Thus, the theoretical models and empirical regularities derived from cross-national studies may not apply in the United Kingdom. And the only study conducted specifically on the British electorate did not examine a variety of alternative explanations. In sum, while previous studies support a variety of conjectures about why the British vary in their support, we lack empirical evidence that identifies which, if any, of these hypotheses apply to the British context. This study attempts to provide evidence of this sort. 5

The remainder of the paper is organized as follows. The next section reviews the relevant literature and identifies specific hypotheses relevant to the data available for analysis. Section two describes the data and measures used to test these hypotheses. We analyze Eurobarometer 53, which was conducted in Spring 2000. This is the most recent Eurobarometer that includes relevant questions for analysis. The third section discusses the results of the analysis and their implications for understanding public opinion for the Euro in the UK. I. Literature Review: Explaining Individual-Level Attitudes in Britain Towards the Euro Previous research on public support for the Euro offers three types of hypotheses. First, several studies have examined cross-national differences in aggregate national-level support for the Euro (e.g., Kaltenthaler and Anderson 2001). These studies do not generate inferences about indvidual-level variation in support and thus we ignore them here. Second, many of these aggregate-level studies also examine cross-temporal variation in support. For example, Banducci, Karp, and Loedel (2002) show that the EU national publics vary in support with the level of inflation. This research could explain why British citizens vary in support over time. However, because we do not have survey data from more than one time period, we cannot test these hypotheses in the current analysis. Consequently, we focus or review on hypotheses of the third type: hypotheses that explain variation across citizens at the individual-level, independent of time. a. Resource-Based Distributional Consequences of a Single Currency 6

Gabel (2000), Kaltenthaler and Anderson (1999), Scheve (1999), and Banducci, Karp, and Loedel (2002) all identify citizens who are likely economic beneficiaries or victims of the adoption of a single currency. Due to data limitations, we cannot examine several interesting hypotheses regarding occupational sector and trade sensitivity. However, we can examine one central argument in these studies: that a single currency works to the advantage of those with high levels of human and financial capital and to the detriment of those with low incomes and low human capital. As Scheve (1999) demonstrates theoretically and empirically, support for monetary integration in Europe should be positively related to financial assets and occupational skills. Notably, Scheve (1999) provides evidence that financial assets and the level of education (a proxy for human capital) are positively related to support for a single currency in the British electorate in 1997. b. Class-Based Distributional Consequences of a Single Currency Oatley (1997) argues that the politics of inflation involve a conflict between the interests of labor and the political Left and the interests of capital and political Right. For the Left, monetary policy that accommodates wage pressures and tight labor markets is preferable to monetary policy that supports slack labor markets. The Right prefers monetary policy that preserves price stability and loosens labor markets. Joining the Euro requires a commitment to price stability as the goal of monetary policy, thereby resolving this political conflict in favor of the Right. Thus, we would expect support for the Euro to be higher among British citizens who identify themselves on the Right of the Left-Right dimension than citizens who place themselves on the Left. 7

c. National Identity A variety of recent studies highlight the importance of national identity for understanding public attitudes toward European integration, in general, and a single currency, specifically. As Laffan (1996) argues, the EU is developing political authority over areas that are of enormous salience to citizens sense of national identity. Several scholars have explored whether and how citizens national identity and/or European identity relates to their attitudes on European integration. For example, McLaren (2002) shows that support for European integration is related to citizens perceptions of cultural threat. Carey (2002) finds that, for those who identify with strongly with their country, pride in nationality is a strong predictor of support for European integration. Research on support for a single currency discovers similar relationships. National identities involve a national economy with boundaries, which gives the currency an important symbolic value (Lafan 1996: 85; Smith 1991: 14). M ller-peters (1998) distinguishes between patriotic and nationalistic attachments to country. The key distinction between these aspects of identity is that patriotism involves exclusively positive emotional attachment and pride while nationalism rests on negative evaluations of other nationalities. She finds that pride in a national currency is closer to nationalism than patriotism and consequently a strong national identity should be associated with opposition to the Euro. This is particularly true for the British, where nationalism is by far the most important explanation for anti-euro sentiment (M ller-peters 1998). 1 She also finds that support for the single currency is higher among citizens with European level patriotism. However, it is important to note that this study also finds that the general model of national attachments is not a good fit for the United Kingdom. 1 Kaltenthaler and Anderson (1999) also found that one s national pride is negatively related to support for the Euro in a cross-national study. 8

Routh and Burgoyne (1998) develop a theory of national identity and support for the Euro exclusively in the British public. They were interested in identifying the psychological bases for adherence to the status quo in terms of monetary policy versus adoption of the Euro. They hypothesized that one s sense of national identity is a strong determinant of how much value one puts on the status quo. National identity is separated into two components: cultural and instrumental. Cultural identity captures symbolic components of nationhood such as a currency. Instrumental identity involves the tangible benefits of citizenship, such as the quality of government services and policy. The question of support for the Euro thus amounts to a comparison of the instrumental and cultural value of the UK government controlling monetary policy versus the economic benefits and costs of the Euro. Or, as Leonard (2001) puts it, are the economic benefits worth the political price in terms of sovereignty and loss of democratic control? Using pride in the United Kingdom as an indicator of cultural identity and evaluations of the quality of public services as an indicator of instrumental identity, Routh and Burgoyne (1998) find that only pride has a direct effect on support for the single currency. Instrumental identity has an indirect impact on support by tempering one s assessments of the economic benefits of the Euro. But, as we discuss in the next section, there are alternative ways to operationalize these types of identity that may lead to different empirical findings. d. Knowledge and Information about the Euro A variety of studies have noted that the level of public sophistication regarding European issues, generally, and the single currency, specifically, is low. This matters for support because uninformed citizens tend to have less stable attitudes that, consequently, are relatively less 9

structured by systematic factors than informed citizens (Zaller 1992; Alvarez and Brehm 1995). As a result, we would expect that the variability in survey responses to questions about the Euro to be greater among citizens who are poorly informed about the Euro those who are well informed about the Euro (Atkinson 2001). Also, previous studies show that support for the Euro increases with the level of public information and saliency regarding the Euro (Whitely 2001; Huhne 2001). One reason for this may be that opposition is based on general fear or ignorance of the single currency and greater knowledge reduces these concerns. e. General Attitudes Toward the EU Banducci, Karp, and Loedel (2002) argue that citizens attitudes toward the EU affect how they view the single currency. A positive experience or view of the EU provides citizens with greater confidence in the conduct of monetary policy at the EU level than a negative appraisal of membership in the EU. This is roughly consistent with the general decision model defined by Routh and Burgoyne (1998), whereby citizens are comparing the status quo to the Euro. British citizens experience with the EU should affect how they appraise the value of adopting the Euro. 10

Data and Measurement The literature reviewed above identifies a variety of hypotheses about why citizens vary in their support for the Euro. We examine these hypotheses with data collected by the Eurobarometer in Spring 2000. 2 We chose this data because it is the most recent Eurobarometer survey that includes a question about support for a single currency asked in the United Kingdom. However, the data have limitations that will constrain how we test some of the hypotheses. Below we describe the specific hypotheses and their operationalization. a. Human and Financial Capital We expect support for the single currency to increase with one s income and one s education. To test this hypothesis we created two variables: education and income. Education is measured by the age the respondent finished her full-time education. 3 Income is monthly reported income, separated into eleven categories in ascending order. 4 b. Left-Right Ideology We expect support for the single currency to be greater among those on the political Right than among those on the political Left. To test this, we use citizen s self-reported position on a 2 Hartung, Harald. Eurobarometer 53: Racism, Information Society, General Services, and Food Labeling, April-May 2000 [computer file]. 3 rd ICPSR version. Brussels, Belgium: INRA (Europe) (producer), 2000. Ann Arbor, MI: Inter-University Consortium for Political and Social Research (distributors), 2002. 3 If the respondent indicated she was still studying, we assigned her the number of years corresponding to her age. 4 less than 240 Pounds Sterling, 240-319, 320-399, 400-479, 480-599, 600-729, 730-829, 830-999, 1000-1149, 1150-1664, 1665-2000, 2000+) 11

ten point Left-Right scale, where (1) represents the Left and (10) represents the Right. 5 We created three dummy variables from this rang: Left (1-4), Center (5-6), Right (7-10). 6 c. National Identity The previous literature supports several hypotheses related to national identity. First, we expect citizens with strong national identity and pride in their nation to oppose the single currency while citizens who have a European identity to support the single currency. The Eurobarometer includes two questions that tap these sentiments: Identity: In the near future, do you see yourself as? (British only; British and European; European and British; European only). National Pride: Would you say that you are very proud, fairly proud, not very proud, not at all proud to be British? (3-very proud; 2-fairly proud; 1- not very proud; 0-not at all proud). We created dummy variables for each of the response categories to the Identity question. We expect respondents who see themselves as British only to express less support for the Euro than those who see themselves as European only in the future. We expect National Pride to be negatively related to support for the single currency. We also specified an interaction term consistent with Carey (2002) in his treatment of national identity and support for European integration. He posited that national pride is primarily relevant to those with strong national identity. Thus, he interacted the dummy variable indicating those who see themselves as exclusively British with the National Pride variable. This new 5 The question reads: In political matters people talk of the left and the right. How would you place your views on this scale? 1-Left, 10-Right. 6 146 respondents either refused to answer or answered don t know. We deleted these respondents from the analysis involving this variable. 12

variable, National Pride*National Identity should be negatively related to support for the single currency. Routh and Burgoyne (1998) make a distinction between cultural national identity, captured by the variables described above, and instrumental national identity. Instrumental national attachments reflect citizens considerations of the quality of domestic institutions. Thus, we would expect support for the Euro to be inversely related to citizens satisfaction with the way democracy works in their nation. To capture instrumental attachments, we use the respondent s level of satisfaction with the way democracy works in the United Kindgom: Satisfaction with Democracy (UK): On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied with the way democracy works in the United Kingdom? 0- not at all satisfied; 1-not very satisfied; 2-fairly satisfied; 3-very satisfied. 7 Finally, we expect that citizens, in comparing the status quo with the Euro will take into account their judgments of the quality of EU institutions. This is a component of instrumental attachments to the EU that should be relevant to citizens consideration of changing authority over monetary policy to the EU level. The question is worded the same as the question for the national level: Satisfaction with Democracy (EU): And how about the way democracy works in the European Union? (0-not at all satisfied; 1-not very satisfied; 2-fairly satisfied; 3-very satisfied) 8 d. Information We expect information to have two distinct effects on support for the single currency in the UK. First, we expect well-informed citizens to have less variable responses to the survey question than those who are poorly informed about the Euro. Put differently, we expect 7 The 79 responses of don t know were coded as 1.5. 13

responses by informed citizens to be structured more by systematic factors such as those hypothesized above than citizens who are poorly informed. To test this claim, we will estimate the error variance as a function of the level information of the respondent. Second, we expect those who are more informed to be more supportive of the single currency. To measure a respondent s level of information, we use the following survey question: Informed: How well informed do you feel about the single European currency, that is the euro? (0-not at all informed; 1-not very well informed; 2-well informed; 3-very well informed) e. Attitudes Toward the EU We expect citizens appraisals of the single currency to reflect their evaluation of their membership in the EU. To test this, we include two variables: EU Benefit (personal): Do you think you, yourself, have more advantages or more disadvantages from the United Kingdom being a member of the European Union? (0-many more disadvantages; 1-more disadvantages; 2-as many advantages as disadvantages, or don t know 9 ; 3-more advantages; 4-many more advantages) EU Benefit (national): Taking everything into consideration, would you say that the United Kingdom has on balance benefited or not from being a member of the European Union? (0-not benefited; 1-don t know; 2-benefitted) We chose these variables because they capture both sociotropic and individual considerations of the value of EU membership. Presumably, citizens who consider the EU beneficial for them personally or for their country are more favorably predisposed toward granting the EU authority over new policies, such as monetary policy. 8 The 304 responses of don t know were coded as 1.5. 9 183 observations. 14

f. Control Variables Previous studies identify age and gender as related to support for a single currency. Given that these characteristics may be related consistently to independent variables of interest, such as education or level of information, we want to control for these factors in testing the hypotheses. We do not report the coefficients for these variables but they are available upon request. g. Dependent Variable To measure support for the single currency, we used the following question: Support for Single Currency: There has to be one single currency, the euro, replacing the British Pound and all other national currencies in the member states of the European Union. (0- Against, 1-For.) 10 Note that this question allowed responses of don t know, of which there were 194. 1232 respondents remained after deletion of missing data for other relevant variables. We have also estimated the same models presented here but with the inclusion of these responses of don t know as an intermediate category in an ordered probit. The results support similar inferences. Table 1 presents some descriptive statistics on the variables. 10 The 194 responses of don t know were deleted from the analysis. 15

Table 1. Descriptive Statistics Variable Mean Standard Deviation Support for Single Currency.278.448 EU Benefit (personal) 1.97.87 EU Benefit (national).92.86 Satisfaction with Democracy (UK) 1.53.82 Satisfaction with Democracy (EU) 1.38.74 Identity 1.48.73 National Pride 1.43.62 Informed 1.06.80 Education 17.00 3.41 Income 6.97 3.11 h. Methodology To test the hypothesized relationships, we estimated a series of probit and heteroskedastic probit models. The probit model is appropriate for the dichotomous dependent variable. But, we do not expect a common error variance across observations. Due to their level of information regarding the Euro, we expect respondents to differ in the variability or ambivalence of their opinions on the Euro. Consequently, we would expect the error variance to decrease with the level of information about the single currency. If not accounted for, the resulting heteroskedasticity causes biased and inconsistent estimates. There are two remedies to this problem in this data setting (see Beck and Tucker 1996). First, following Alvarez and Brehm (1995) and Greene (1997: 649), w model a multiplicative heteroskedastic error variance (Harvey 1976) where this variance is a function of the level of information of the respondent. We expect a negative coefficient on this parameter, indicating that as information increases the size of the 16

error variance decreases. Second, I estimate Huber (1967) robust standard errors for all parameters, which accounts for any heteroskedasticity not captured in the error variance model. Results Tables 2 and 3 report the results of several models. Models 1-3 examine the effect of national identity under different specifications. Models 4 and 5 report the results for tests of the effects of income and ideology. These are reported separately because these variables had significant numbers of missing values, with reduces the sample size. Table 2. Models of Support for a Single Currency Model 1 Model 2 Model 3 EU Benefit (personal).201 (.063).203 (.064).203 (.064) EU Benefit (national).256 (.061).258 (.062).260 (.062) Satisfaction with Democracy (EU).325 (.074).337 (.076).336 (.076) Satisfaction with Democracy (UK) -.106 (.060) -.108 (.060) -.108 (.060) National Pride -.005 (.080) -.014 (.084) - National Identity -.433 (.346) -.395 (.356) -.500 (.091) National Pride*National Identity -.035 (.131) -.041 (.136) - European + National Identity -.129 (.201).132 (.191) European Identity -.163 (.214).163 (.213) Education.010 (.011).011 (.011).011 (.011) Informed.232 (.053).224 (.054).224 (.054) Constant -1.32 (.376) -1.378 (.397) -1.472 (.332) Error Variance Model Informed -.135 (.070) -.124 (.073) -.125 (.073) Heteroskedasticity Test Likelihood Ratio Test 3.74* 2.90** 2.92*** N 1038 1038 1038 *prob. > _ 2 =.05 ;** prob. > _ 2 =.09; *** prob. > _ 2 =.09 17

Table 3. Models of Support for a Single Currency (continued) Model 4 Model 5 EU Benefit (personal).279 (.116).210 (.073) EU Benefit (national).310 (.118).240 (.066) Satisfaction with Democracy (EU).290 (.106).339 (.081) Satisfaction with Democracy (UK) -.005 (.101) -.114 (.064) National Identity -.413 (.154) -.560 (.102) European + National Identity -.039 (.303).177 (.207) European Identity.750 (.467) -.022 (.223) Education -.010 (.018).009 (.013) Informed.104 (.098).256 (.060) Income -.001 (.021) - Left -.163 (.105) Right - -.227 (.130) Constant -1.11 (.534) -1.575 (.394) Error Variance Model Informed -.093 (.129) -.104 (.077) Heteroskedasticity Test Likelihood Ratio (χ2).053 1.83 N 400 893 The results relating to the significance of the variables can be summarized as follows. First, information has the expected effects on support. The more information citizens have about the Euro, the more they are likely to support British membership of the single currency. Figure 1 plots the relationship between the probability of supporting a single currency and the level of information the respondent reports regarding the Euro. 11 A change in a respondent s level of information from lowest to highest is associated with a 100 percent increase in the probability of 18

support. A one standard-deviation increase (.80) from the mean value of informed (1.06) is associated with about a seven percent increase in the probability of support. Figure 1: The Effect of Information on Support 0.55 Probability of Support 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0 1 2 3 Level of Information Information also appears to have the expected effect on the variability in responses. The error variance model indicates that as a respondent s level of information about the Euro increases, her error variance decreases. In Models 1-3 this parameter is significant at the.10 level or better. The Chi-squared test for homoskedasticity is also rejected at the.10 level for all three models. Note that the effect of information on the error variance influences the impact of the independent variables in the model of support for the Euro. The basic point is that the independent variables have a greater impact on support among the well-informed respondents than among the poorly informed respondents. This is captured in Figure 1, as increases in informedness affect support through both channels. But the interactive effect of information is more apparent in subsequent figures for other independent variables. 11 These predicted values are calculated at the mean for all variables except National Identity, which is set at zero 19

Second, the results are mixed regarding the importance of economic calculations. On the one hand, objective measures of individual economic benefit/loss from British membership of the single currency do not affect individual attitudes towards the single currency. There is no relationship between the level of personal socio-economic resources of a citizen (their educational level or their income level), which determine whether an individual is likely to gain or lose from British membership of a single currency, and whether the citizen then supports or opposes British membership of the single currency. On the other hand, perceptions of economic benefits from the European Union as a whole do affect individual attitudes towards British membership of the Euro. If a citizen perceives that he/she personally benefits from EU membership, or if a citizen perceives that Britain benefits from EU membership, then he/she will be more likely to support British membership of the Euro than a citizen with opposite perceptions. Figure 2 presents the relationship between evaluations of personal economic benefit from the EU and support for the Euro. 12 The figure also shows how this affect varies depending on the information level of the respondent. These personal evaluations have a dramatic effect on support, particularly for those who are well informed. For well-informed respondents, a one-standard deviation (.87) increase from the mean evaluation (1.97) is associated with about a ten percent higher probability of support. (the modal response). 12 These predicted values are calculated at the mean for all variables except National Identity, which is set at zero (the modal response). 20

0.8 Figure 2: The Effect of Evaluations of Personal Benefit on Support Probablity of Support 0.7 0.6 0.5 0.4 0.3 0.2 0.1 At Lowest Level of Information At Highest Level of information At Mean Level of Information 0 0 1 2 3 4 Evaluatoin of Personal Benefit Third, some political factors matter more than others. Where the EU is concerned, there is strong and positive relationship between whether a person believes that the EU is democratic and whether he/she supports British membership of the Euro. Put another way, playing up the existence of a democratic deficit in the EU is an effective strategy for the anti-euro campaign. However, identification with Europe does not affect support for the single currency. Figure 3 shows the relationship between support and satisfaction with democracy in the EU. 13 Satisfaction with democracy in the EU has a sizable impact on support. For a well-informed respondent, a one standard deviation increase in satisfaction (.74) from the mean level (1.38) is associated with more than a ten percent increase in the probability of supporting a single currency. And, a change from least to highest level of satisfaction is associated with a forty percent increase in support, for a respondent at the mean level of information. 13 These predicted values are calculated at the mean for all variables except National Identity, which is set at zero (the modal response). 21

Figure 3: The Effect of Satisfaction with EU Democracy and Support Probability of Support 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 Satisfaction with EU Democracy At Mean Level of Information At Lowest Level of Information At highest level of information Where political attitudes to Britain are concerned, national identity (that is, only seeing oneself as British in the near future) is negatively associated with support for British membership of the Euro. However, whether a person is proud of Britain does not make a difference, and we do not find support for Carey s expected interaction between national identity and pride. Table 4 shows the predicted probabilities for changes in national identity under different assumptions about a respondent s level of information. 14 In general, a change in a respondent s identity to that of feeling British exclusively is associated with about a fifty percent decrease in the likelihood of supporting the Euro. Table 4. Effect of National Identity on Support Not British exclusively British Exclusively Lowest information level.309.143 Mean information level.251.121 Highest information level.498.234 14 These predicted values are calculated at the mean for all variables. 22

Support is also related to satisfaction with democracy in the UK in the expected way. As respondents levels of satisfaction increase, their likelihood of supporting the single currency decreases. Finally, political ideology does matter, but in the opposite direction to the relationship found in other member states. In Britain, voters on the Left are more supportive of British membership of the single currency than voters on the Right. Conclusions Before discussing the possible implications of our findings, we want to note that our data limits how strongly we can infer referendum voting behavior from our results. Ideally, we would use survey data including a question specific to the referendum and with a set of socio-economic and political questions tailored to the theoretical expectations we described in the literature review. We would also prefer to use some time-series data to capture dynamic effects that we cannot directly examine here. However, our data do allow us to advance our understanding of why British citizens vary in their support for a single currency. We are aware of no study of British individual-level public opinion that directly examines the hypotheses described here with survey data from a more recent year than the data analyzed here. Thus, the analysis provides an advance in our empirical inferences about British public opinion toward the Euro. Our results suggest several possible factors that could influence the outcome of a referendum on the Euro in the United Kingdom. Many of the factors that have substantively significant effects on support are, at least in principle, malleable. Certainly, how well British citizens are informed about the single currency can increase dramatically with an information and political campaign surrounding a referendum. Indeed, Whiteley (2001) shows that the discussion 23

of the Euro in the campaign preceding the 2001 British parliamentary election had a discernable impact, consistent with the results presented here, on both the level of public awareness regarding the Euro and the level of support for the Euro among voters. If the public were informed by a campaign, the results suggest that this would have two distinct effects on support. First, it would generally raise support for the single currency. Second, it would lead to greater divisions in support due to the factors described here (e.g., satisfaction with democracy at the EU and UK levels). Consequently, either side of the campaign could benefit by using the campaign to stress attempt to impact how citizens appraise the benefits of EU membership and satisfaction with the way democracy works at the national and EU levels. By this account, the focus on the lack of democracy at the EU level by opponents of single currency appears well targeted (Curtice 2001). And, this also suggests that proponents of the Euro should resist blaming Brussels when it might be expedient for national political reasons. More concretely, a study by Karp, Banducci, and Bowler (forthcoming) indicates that voters satisfaction with democracy at the EU level varies in systematic ways. In particular, satisfaction increases with the level of net financial benefits from the EU budget. Thus, to the extent other EU member-states would like to effect British participation in the single currency, their decisions about the EU budget could be used to affect the outcome of a referendum. It is also interesting to note that our results indicate that economics matter less at the individual-level than Brown might think. While we do not examine whether all possible distributional consequences of a single currency influence support, the general effects due to income and education do not have the expected effects on British public opinion. Thus, making the case for the benefits of a single currency may not affect support. However, we do find that citizens evaluations of personal and national benefits of EU membership are an important 24

determinant of support. Thus, a pro-euro campaign might want to focus its economic arguments on the benefits derived from the EU. 25

References Alvarez, Michael and John Brehm. 1995. American ambivalence towards abortion policy: Development of a heteroskedastic probit model of competing values. American Journal of Political Science 39(4): 1055-1082. Banducci, Susan, Jeffrey Karp, and Peter Loedel. 2002. The Euro, Economic Interests, and Mulit-Level Governance: Examining Support for the Common Currency. European Journal of Political Research forthcoming. Beck, Nathaniel, and Richard Tucker. 1996, Conflict in space and time: T ime-series-crosssection analysis with a binary dependent variable. Paper presented at the 1996 Annual Meeting of the American Political Science Association, San Francisco, August 29-September 1, 1996. Carey, Sean. 2002. Undivided Loyalties: National Identity as Obstacle to European Integration? European Union Politics forthcoming. Dalton, Russell and Robert Duval. 1986. The Political Environment and Foreign Policy Opinions: British Attitudes toward European Integration, 1972-1979. British Journal of Political Science 16(1): 113-134. Gabel, Matthew. 2000. Divided Opinion, Common Currency: The Political-Economy of Public Support for EMU. In Barry Eichengreen and Jeffry Frieden, Eds. The Political Economy of European Monetary Unification. Boulder: Westview Press. G rtner, Manfred. 1997. Who Wants the Euro and Why? Economic Explanations of Public Attitudes Towards a Single European Currency. Public Choice : 93: 487-510. Greene, William H. 1997. Econometric analysis, 3rd ed (Upper Saddle River, N.J.: Prentice Hall). Harvey, Andrew. 1976. Estimating regression models with multiplicative heterskedasticity. Econometrica 44: 461-465. Hix, Simon. 2000. Britain, the EU and the Euro. In Patrick Dunleavy, Andrew Gamble, Ian Holliday and Gillian Peele (eds) Developments in British Politics 6, London: Palgrave. Huber, Peter J., 1967, The behavior of maximum likelihood estimates under non-standard conditions, in: Proceedings of the Fifth Annual Berkeley Symposium on Mathematical Statistics and Probability, Vol. 1 (Berkeley: University of California Press, 221 233). Huhne, Christopher (2001) Winning on Economics, in Mark Leonard and Tom Arbuthnott, Eds. 2001. Winning the Euro Referendum. London: the Foreign Policy Centre. Kaltenthaler, Karl and Christopher Anderson. 2001. Europeans and their money: Explaining Public Support for the Common Currency. European Journal of Political Science 40: 139-170. 26

Kaltenthaler, Karl and Christopher Anderson. 1999. Europeans and Their Money: Explaining Public Support for the Common European Currency. Working Paper, Department of Political Science, SUNY-Binghamton. Karp, Jeffrey, Susan Banducci, and Shaun Bowler. Forthcoming. To Know it is to Love it? Satisfaction with Democracy in the European Union. Comparative Political Studies. Kokkinaki, Flora. 1998. Attitudes Towards European Monetary Union in Greece: Antecedents, Strength and Consequences. Journal of Economic Psychology 19: 775-796. Laffan, Brigid. 1996. The Politics of Identity and Political Order in Europe. Journal of Common Market Studies 34 (1): 82-102. Leonard, Mark (2001) Wining on Politics, in Mark Leonard and Tom Arbuthnott, Eds. 2001. Winning the Euro Referendum. London: the Foreign Policy Centre. Leonard, Mark and Tom Arbuthnott, Eds. 2001. Winning the Euro Referendum. London: the Foreign Policy Centre. McLaren, Laura. 2002. Public Support for the European Union: Cost/Benefit Analysis or Perceived Cultural Threat? Journal of Politics 84 (2): 551-566. M ller-peters, Anke. 1998. The Significance of National Pride and National Identity to the Attitude Toward the Single European Currency: A Europe-Wide Comparison. Journal of Economic Psychology 19: 701-719. Routh, David and Carole Burgoyne. 1998. Being of Two Minds About a Single Currency: A UK Perspective on the Euro. Journal of Economic Psychology 19: 741-754. Scheve, Kenneth. 1999. European Economic Integration and Electoral Politics in France and Great Britain. Working Paper, Department of Political Science, Yale University. Smith, Andrews. 1991. National Identity. Harmondsworth: Penguin. Van Everdingen, Yvonne and W. Fred van Raaij. 1998. The Dutch People and the Euro: A Structural Equations Analysis Relating National Identity and Economic Expectations to Attitudes Towards the Euro. Journal of Economic Psychology 19: 721-740. Zaller, John. 1992. The nature and origins of mass opinion. Cambridge: Cambridge University Press. 27